Adaptive Collapsing on Bounding Volume Hierarchies for Ray-Tracing

Abstract

Ray tracing is a computationally intensive process, several tree data structures and heuristics have been developed to optimize it. This paper presents a new heuristic in the area, based on collapsing some nodes in order to achieve a smaller expected number of node-tests. Two ways of using this heuristic in Bounding Volume Hierarchies are presented as well as the cost-model used to drive the heuristic development and measure it s efficiency. Some procedures on integrating this heuristic with other optimizations are also discussedRay tracing is a computationally intensive process, several tree data structures and heuristics have been developed to optimize it. This paper presents a new heuristic in the area, based on collapsing some nodes in order to achieve a smaller expected number of node-tests. Two ways of using this heuristic in Bounding Volume Hierarchies are presented as well as the cost-model used to drive the heuristic development and measure it s efficiency. Some procedures on integrating this heuristic with other optimizations are also discussed